The art in front of artificial intelligence

Approximately three years after the popularization of prompt-based technologies such as ChatGPT, Dall-E, Midjourney, and Sora, what are the interactions between art and artificial intelligence (AI)? And what is the relationship between the art world and the contemporary technological landscape? Among both laypeople and specialists, there is a common understanding that we have entered a new era of artificial intelligence, marked by much speculation and limited knowledge about its full potential.

When we talk about AI, we generally refer to models developed from foundation infrastructures built by a few U.S.-based technology corporations. These companies develop the bulk of the systems—which are quite expensive and resource-intensive—that are later sold to smaller developers, so they don’t have to start their projects from scratch.

In his book "Art and Cosmotechnics", philosopher Yuk Hui warns of the risk of considering AI as neutral and universal. According to him, AI is not merely a technical project but also a metaphysical one, based on a Cartesian view that reduces human intelligence to something measurable and reproducible by machines. Hui suggests that AI can be developed differently, inspired by non-Western cosmologies, to enhance human creativity and promote interconnectedness, rather than being limited to the pursuit of efficiency.

Talking about the history of art is also discussing the development of technologies and media

— Livia Benedetti

Among art's forays into the field of digital technology, about five years ago, NFTs and the metaverse were heralded with great exaggeration as new paradigms, only to be sidelined shortly thereafter. On the other hand, artificial intelligence has proliferated in an unavoidable manner. In a time of hyper-consumption of images on the internet, deepfake content generated by autonomous systems quickly blended indistinguishably into the stream of images — with more people uploading files that are nearly impossible to discern as “real” or “fake”. This duality has not existed for a long time, as images have been manipulated practically since the creation of photography. From the 19th-century analog "spirit photographs", which forged the capture of paranormal manifestations through light exposure, to software like Photoshop, used for over thirty years, we have long had reasons to mistrust what we see.

The fact is that, at this moment, the general notion that AI mimics reality and human creation triggers feelings of apprehension and repulsion widely documented on social media. Despite being an extremely advanced technology, a text that “seems to have been written by ChatGPT” today is synonymous with a mediocre, poor, or unoriginal text. Recently, we have seen the emergence of the first art pieces generated by prompts in AIs like Midjourney. Initially, these works dazzled with the advancement of technology, but within a few months, as access expanded, they quickly became banal.

One of the most intriguing figures at the intersection of art and artificial intelligence is artist and composer Holly Herndon, who, alongside her partner Mat Dryhurst, has been exploring the use of machine learning in art, particularly in music. In 2019, on her album PROTO, she introduced Spawn, a synthetic voice algorithm that interacted with her in the compositions. Herndon views AI not just as a tool but as a potential creative partner. She recently developed Holly +, an algorithm trained with her voice that can perform by singing songs by others without her presence. Could these creations be considered collaborations with Herndon? Through these experiments, she challenges and expands notions of artistic authorship and uniqueness, while also elevating the longstanding debate about artists' appropriation and consent to new dimensions. Expanding this discussion, Dryhurst and Herndon created a website called Have I Been Trained?, where artists can find out if their works are being used to train AIs and proceed to prohibit the use of that data if desired.

Another significant AI-generated artwork is "Unsupervised" (2022) by Refik Anadol, a public and media phenomenon showcased at MoMA in New York. It features a large LED screen installed in the lobby, where moving images are generated through an algorithm that interprets data from the museum's collection. This visual hypnosis was sarcastically dubbed “MoMA’s Glorified Lava Lamp” by art critic Jerry Saltz in a review for Vulture magazine, indicating a certain skepticism about artistic experiments with artificial intelligence. With this specific work, Anadol has boosted his career as an artist and conceived Dataland, the first museum dedicated to AI-created art, which is set to open this year in Los Angeles.

In the face of the social dominance of digital technologies, recent years have seen a growing interest in craftsmanship

— Livia Benedetti

To talk about the history of art is also to talk about the history of the development of technologies and media. The emergence of these, from the most rudimentary to the most digitally advanced, has always impacted artistic creation: oil paint gave mobility to Impressionist painters, allowing them to work better outdoors; the advent of photography influenced the understanding of painting; the use of computers in place of typewriters enabled copy-paste and new forms of writing, as observed by Kenneth Goldsmith; there are numerous examples.

In art, there have been many moments when the values associated with industrial production began to be celebrated at the expense of artisanal techniques and traditions. We can think of several movements that occurred at different times, such as Futurism, the Bauhaus, and American Minimalism. Currently, it wouldn't be an exaggeration to say that the direction is reversed. In the face of the social dominance of digital technologies, in recent years there has been a growing search for crafts, handiwork, ancestral languages, and cosmologies. An example of this is the meteoric appreciation of Indigenous culture in the Brazilian art market. This is, in part, a response to the dystopian times intensified by the internet and digital oligarchies. If one of the greatest fears today is the replacement of sectors of human activity by AI, the realm of art tends not to be one of them.

Livia Benedetti is an art curator, researcher, and co-founder of the platform aarea.co. She pursued her master's in art mediation at Université Paris VIII and Visual Arts at the University of São Paulo. Since 2007, she has been involved in curatorial and educational projects in cultural institutions such as Sesc, Pivô, and the São Paulo Cultural Center. In 2016, she was an associate curator of the Danish Biennale FotoBiennale at the Brandts Museum, and in 2019, she participated in the Emerging Professionals program at Para Site in Hong Kong. As a writer, she has been publishing her texts in magazines, books, and art catalogs.